Introduction

Kenneth N. Waltz’s Theory of International Politics profoundly affected international relations theory. Structural realism as developed by Waltz argues that the anarchic system and the distribution of capabilities are powerful constraints and inducements which produce “sameness” in the behavior of states. For Waltz, international relations is anarchic and not hierarchical, populated by functionally similar units, and the structure of the international system or polarity varies based on the distribution of capabilities. The anarchic nature of the international system, and the assumptions that states “at a minimum, seek their own preservation” and are socialized to imitate each other, allows Waltz to explain recurring international patterns and outcomes such as balances of power, war proneness of different distributions of power, and recurrent alliance formation (1979:118; for realist theories of foreign policy, see the literature on classical realism such as Gulick 1955; Wolfers 1962; Morgenthau 1963; Thucydides 1982; and neoclassical realism including Friedberg 1988; Snyder 1991; Wohlforth 1993; Christensen 1996; Elman 1996; Rose 1998; Schweller 2006; Rathbun 2008; Lobell et al. 2009). Waltz’s structural realism influenced many of the major debates in the field in the 1980s and 1990s including neoliberal institutionalism (Keohane 1984; Oye 1986; Baldwin 1993; Ruggie 1993); the agent–structure debate (Wendt 1987; Dessler 1989); the significance of non-state actors (Krasner 1983); and more recently, the new international hierarchy studies (Lake 2003; 2009; Hobson and Sharman 2005) and the degenerative research program controversy (Vasquez 1997; Legro and Moravcsik 1999). Criticism and dissatisfaction with Waltz’s structural realism (Ruggie 1983; Ashley 1986; Keohane 1986; Buzan et al. 1993) fueled the constructivist, cultural, ideational, and Innenpolitik research agendas (Wendt 1992; Goldstein and Keohane 1993; Rosecrance and Stein 1993; Katzenstein 1996; Adler and Barnett 1998; Guzzini 1998).

Derived from Waltz’s structural realism, structural realist theorists can be divided into two competing versions with competing assumptions and policy prescriptions: offensive realism and defensive realism (for reviews of this literature see Lynn-Jones and Miller 1995; Frankel 1996; Miller 1996; Brooks 1997; Taliaferro 2000/01; James 2002; Walt 2002; Schweller 2003; Nexon 2009; the terms “aggressive” and “defensive” realism originally appeared in Jack Snyder’s Myths of Empire). One distinction between these two versions of realism is the role of the anarchic international system and whether it encourages states to maximize their security or to maximize their power and influence. A second distinction is whether conquest and expansion pay, and more generally, the cause of pathological state behavior including overexpansion, selfencirclement, and overextension. A final distinction is whether states are primarily revisionist in their intentions, or at least assumed to be, or whether states are primarily motivated by security-seeking behavior.

For offensive realists security is scarce. The anarchic nature of the international system compels states to maximize their share of world power and to seek superiority, rather than equality, in order to make themselves more secure and thereby increase their odds of survival (Gilpin 1981; Liberman 1996; Schweller 1996; Labs 1997; Zakaria 1998; Mearsheimer 2001; Elman 2004). The ultimate goal of every major power is to become the hegemon. The rationale is that the more power and the stronger the state, the less likely it will be a target, since weaker powers will be dissuaded from challenging it. John Mearsheimer is clear that “states quickly understand that the best way to ensure their survival is to be the most powerful state in the system” (2001:33). Uncertainty about intentions of other states combined with the anarchical nature of the international system compels great powers to adopt competitive, offensive, and expansionist policies whenever the benefits exceed the costs. Specifically, since intentions are never clear and a state might become more aggressive in the future, all of the major powers adopt a worst-case scenario and therefore increase their power through expansion which leads to high levels of competition. Moreover, for offensive realists, offensive actions often succeed and conquest often pays.

For defensive or positional realists (Joseph Grieco coined the term “defensive positionalists” in Cooperation Among Nations), security is plentiful. Major powers seek to maximize their security by preserving the existing balance of power through mostly defensive strategies (Jervis 1978; Waltz 1979; Posen 1984; Walt 1987; Grieco 1990; Snyder 1991; Glaser 1994/5; Layne 1997; Van Evera 1999). Defensive realists maintain that the international system encourages states to pursue moderate and restrained behavior to ensure their survival and safety, and provides incentives for expansion in only a few select instances. The rationale is that aggression, competition, and expansion to maximize power through primacy and preponderance are unproductive because they will provoke the security dilemma and counterbalancing behavior, and thereby thwart the state’s effort to increase its security. As Christopher Layne concisely notes, “states balance against hegemons” (1993:87). For defensive realists, since the international system rarely provides incentives for expansion, “structural modifiers,” including the offense–defense military balance and geography, and domestic and unit-level pathologies such as elite beliefs, perceptions, and logrolled imperial coalitions, explain overexpansion, underbalancing, self-encirclement, and overextension (Taliaferro 2000/01).

These two competing versions of realism raise several questions: What is the implication of anarchy and does the anarchical structure of the international system encourage states to maximize power or to maximize security? Is security in the international system scarce or plentiful? Does the international system encourage states to engage in offshore balancing, selective engagement, or primacy (Layne 1993; Posen and Ross 1996/97; Art 2003; Posen 2003; Walt 2005; Brooks and Wohlforth 2008)? Does the international system encourage all great powers to act aggressively toward each other, even status quo states? How much power do states want and how much security do they require? Do conquest and territorial expansion pay or are they counterproductive? And how do offensive and defensive realists explain cases of self-defeating outcomes such as overextension, self-encirclement, and overstretch?

Offensive Realism and Maximizing Power

Classical realists (such as Thucydides, E.H. Carr, Arnold Wolfers, and Hans Morgenthau) and offensive realists share the assumption that states seek to maximize power – that states are relentless seekers of power and influence. Specifically, for classical realists “nations expand their political interests abroad when their relative power increases” (Zakaria 1998:19). Thucydides’ assertion in “The Melian Dialogue” is that “the strong do what they can and the weak suffer what they must.” For Morgenthau and other classical realists, the cause of conflict, aggression, expansion, and wars is human lust for power and hence states are hardwired to behave aggressively toward one another. The import is a never-ending struggle among states due to the natural human urge to dominate others.

For offensive structural realists, the anarchic international system and the distribution of power, and not human nature, is the invisible hand that “shapes and shoves” all great powers to maximize power and influence despite domestic or unit-level differences. The international system creates powerful incentives for states to look for opportunities to gain power and influence at the expense of rival states. As Robert Gilpin argues, “as the power of a state increases, it seeks to extend its territorial control, its political influence, and/or its domination of the international economy” (1981:106).

For offensive realists, states maximize power, influence, and wealth (which is the foundation for military power) to become more secure in a world of anarchy; the best way for a state to increase its odds of survival is to become the most powerful state. Simply, a state with more power is more secure than a state with less power. For offensive realists, it is the structure of the anarchic international system that strongly encourages states in their unrelenting pursuit to maximize power with the ultimate goal of becoming a global hegemon, though for Mearsheimer, in contrast to Gilpin, only regional hegemons, such as the US, are possible due to geographical constraints such as the stopping power of water (for a critique of Mearsheimer’s version of offensive realism, see Snyder 2002; Lemke 2004).

For offensive realists, expansion entails aggressive foreign economic, political, and military policies to alter the balance of power; to take advantage of opportunities to gain more power; to gain power at the expense of other states; and to weaken potential challengers through preventive wars or “delaying tactics” to slow their ascent (Organski 1968). As Mearsheimer tells us, only a misguided state would believe that it has an “appropriate amount” of power and pass up opportunities to be the regional hegemon in the system. For if a state does not try to maximize its influence and selects to forgo an opportunity to expand, other powers will take advantage of the opportunity (Zakaria 1998). Thus, a major power does not strive to be an equal among its great-power peers but to be the most powerful – the hegemon. The goal of states is to maximize power and states will always compete with each other for power.

In contrast to Gilpin and other hegemonic offensive realists, for Mearsheimer, geography, and especially the stopping power of water, means that it is impossible for any state to achieve global hegemony. The best a major power can achieve is to become a regional hegemon, that is, the only great power in its part of the globe, and possibly control another region that is nearby and accessible over land. States that achieve regional hegemony seek to prevent great powers in other regions from repeating their accomplishment – they check aspiring or potential hegemons in other locales because they fear that a rival great power that dominates its own region will be an especially powerful foe that can cause trouble in the fearful great power’s locale. Regional hegemons prefer that there be at least two great powers located in other regions in order to check each other’s power.

Fareed Zakaria’s theory of state-centered realism presents a slight twist to the offensive realist argument. He argues that statesmen will expand when they can increase state power and not national power – “nations try to expand their political interests abroad when central decision makers perceive a relative increase in state power” (1998:42). Thus, expansion will allow statesmen to increase the amount of national power they can extract and mobilize.

Status Quo Powers are Rare

For offensive realists, status quo states in the international system are rare and limited to global hegemons. Major powers are rarely satisfied with the current or existing distribution of power. The rationale is that states are never certain of another state’s intentions – whether it can correctly assess its economic and military power, and whether it will use its offensive capabilities in the future to increase its relative power. Specifically, it is difficult for a state to assess how much relative power it must have over its rivals before it is secure and it is difficult to determine how much is enough in the future. Even in the absence of a specific or imminent threat, offensive realists argue that states will maximize power and influence because states cannot be sure when or where the next threat will emerge (Labs 1997). Uncertainty about intentions and fear of miscalculation mean that states always adopt and always prepare for the worst-case scenario when assessing other powers (one of Mearsheimer’s five “bedrock” assumptions about the international system is that all states possess at least some offensive capability). The import is that states always regard each other with fear, mistrust, and suspicion, and moreover, all states think the same way about each other. The result is a constant security competition, even among states that have no reason to compete, and hence the title of Mearsheimer’s book, The Tragedy of Great Power Politics.

For most offensive realists, all major powers have revisionist intentions. Schweller’s balance-of-interests theory, which some (i.e., Frankel 1996; Elman, 1998; Jervis 1999; and Taliaferro 2000/01) have classified as a variant of offensive realism, divides states into status quo and revisionist powers. In contrast to the findings of defensive realists, he argues that revisionist states often bandwagon with the stronger power rather than balance against power (known as jackal bandwagoning). As discussed above, revisionist powers also initiate wars of conquest to add to their power and prestige (Lynn-Jones and Miller 1995:ix–xxi). Status quo states are generally satisfied with what they have and so they tend to balance against threats and seek security. For Schweller, what matters is the aggregate balance between the revisionist and status quo powers, or the balance of interests, which affects the stability and war-proneness of the system.

Conquest Pays

Defensive realists assert that the structure of the international system rarely encourages states to expand in order to increase their security, that conquest is rarely profitable, and that aggression provokes counterbalancing behavior which results in self-encirclement, overextension, and strategic exposure. For defensive realists, this self-defeating behavior is not attributed to systemic pressure but to domestic and unit-level variables. Thus, they find that states are often more secure by maintaining the status quo.

For offensive realists, systemic imperatives push states to expand. Expansion and conquest often make states more secure, can pay huge dividends, and block other states from acquiring additional power. For Gilpin, Mearsheimer, and Zakaria, the quest for greater security encourages states to engage in territorial, political, military, and economic expansion. According to Jack Snyder (1991:21–6), offensive realism stipulates at least four systemic conditions that can promote expansion: (a) when military technology favors the attacker; (b) when states can make significant cumulative additions to their power resources; (c) when relative power is expected to decline and thereby encourages preventive war; (d) when the distribution of power is multipolar, which allows states to defeat opponents piecemeal, and contributes to miscalculations and uncertainty about the actual distribution of relative power.

Offensive realists maintain that conquest is profitable. Mearsheimer claims that states that initiate aggression win the wars 60 percent of the time (“between 1815 and 1980 there were 63 wars and the aggressor won 39 times”; 2001:149). According to Peter Liberman (1996), industrial resources are “cumulative” and can be added readily to existing capability; coercive and repressive subjugators can compel vanquished states to pay a large share of their economic surplus in tax and tribute; modernization vastly increases the economic wealth produced; the richer the country the more that can be taken from it so societal opposition will have to be very costly to make conquest unprofitable; modernization increases the efficiency of foreign coercion because it centralizes control, facilitates the quick deployment of this power over wide regions, and gives hostage societies more to lose from opposition; and relatively low-cost investment in repression prevents most people in modern societies from participating in the collective good of resistance.

For offensive realists, although states are relentless expanders, they are not mindless expanders. States are prudent territorial expanders, especially given the real risk of undermining economic and industrial power, which is the basis for military power. States may forgo opportunities to increase their power because the costs are too high, due to diminishing returns from additional military resources, because it might undermine the economy, or building additional military forces will provoke a rival who can match the increase. Even states with abundant wealth such as the US and UK have selected not to build militaries to conquer Europe or Asia. In response to imperial overstretch arguments (Kennedy 1987), Mearsheimer argues that “little scholarly evidence supports the claim that high levels of defense spending necessarily hurt a great power economy” (2001:149).

For Gilpin (1981), who also argues that major powers are not mindless expanders, a major state will seek to change the international system as long as the expected benefits exceed the expected costs. Furthermore, the state will engage in territorial, political, and economic expansion until the marginal costs of greater expansion equal the marginal benefits. Finally, no state will seek to change the system when it is not profitable.

In response to defensive realists and borrowing from Kenneth Waltz’s arguments about the restraining effect of socialization and learning on state behavior, Eric Labs argues that successful expanders learn from past or previous mistakes. States learn to go about expanding in a manner that draws the least amount of attention from the other major powers (1997:13). Offensive realists acknowledge that states can miscalculate, especially in multipolar distributions where uncertainty about relative power and buck-passing behavior prevail. Moreover, states rarely have complete information about any situation they confront, let alone about the intentions and motivations of rival powers. Yet, even when the outcome fails and the consequence is self-encirclement or overstretch, Mearsheimer asserts that in many cases “a careful analysis shows that these choices for war were a reasonable response to the particular circumstances each state faced” (2001:211). For instance, he maintains that both Nazi Germany and Japan during the 1930s took “a calculated risk” that was unsuccessful (2001:215).

The Most Dangerous Distribution of Power

The distribution of power among the states in the international system also affects the level of fear. Structural realists differentiate between bipolar (two major powers) and multipolar (more than two major powers) distributions of capabilities. Mearsheimer further distinguishes between balanced multipolar and unbalanced multipolar distributions of power. Most offensive and defensive realists (other than hegemonic realists) agree that bipolar systems are more stable and less war prone than multipolar systems, and both are more stable than a unipolar system, though Karl Deutsch and David Singer (1964) counter that multipolar systems are more stable. Offensive and defensive realists disagree on whether multipolar systems, and especially unbalanced multipolar systems (the most unstable distribution), compel a potential regional hegemon to expand to become a regional hegemon or encourage a major power to act with restraint to prevent counterbalancing behavior.

For Mearsheimer the configuration of power that generates the most fear and the greatest security competition among states is a multipolar system that contains a potential hegemon or what he calls an unbalanced multipolar system (several great powers and a potential regional hegemon). As discussed below, geographic factors such as contiguity further heighten the fears of expansion. Pressure to expand is great for a potential regional hegemon because it strives to become a regional hegemon in order to increase its odds of survival, and because of its relative power, it has a good chance of dominating and controlling the other great powers in the region. Pressure to expand is further exacerbated in regions with continental powers that have large land armies – since for Mearsheimer these are the states that have initiated most of the wars of conquest. Thus, a potential regional hegemon will behave more aggressively than a major power facing other major powers in the same locale.

For Mearsheimer, regional hegemons block peer competitors in other regions from achieving regional hegemony. As mentioned above, the rationale is that a peer regional hegemon in another locale might support a rival or cause problems in other locales. Therefore, an extra-regional hegemon will assist other powers in the region to ensure that there are always at least two great powers in the locale.

In balanced multipolar systems, security competition is also high, but less so than in unbalanced multipolar systems. For Mearsheimer, buck-passing is more common in a balanced multipolar system, and even more so, if there are geographic barriers (while balancing is more prevalent in unbalanced multipolar systems and among contiguous states). Where buck-passing rather than balancing is more prevalent, expanders will face less opposition and greater opportunities to expand while the other major powers are engaged in passing the buck and debating who will bear the burden of balancing the aggressor.

A bipolar system is the most stable system and produces the least amount of fear among the great powers (Dale Copeland asserts the opposite; 1996; 2000). In this system, there is usually a rough balance between the major states in the system. Buck-passing cannot occur, and bipolarity discourages miscalculations and is more efficient since balancing occurs through internal mobilization (rather than counterbalancing alliances which can be slow to form).

Geography also contributes to the level of threat. Mearsheimer (like defensive realists who include “structural modifiers” in their arguments, see below) includes the role of geography by differentiating between continental and insular great powers. For Mearsheimer, insular states like the UK and the US are less vulnerable to invasion than continental states. Given the stopping power of water, which makes it difficult for states to project their power over long distances, insular states are more secure and less vulnerable to conquest than continental states. Continental great powers including France, Germany, and Russia/USSR are more likely to make bids for regional hegemony and more likely to face balancing coalitions than geographically distant or insular great powers.

Balancing versus Buck-Passing

For offensive and defensive realists, the choice of balancing and buck-passing against an aggressor is a function of the structure of the international system. A threatened great power in a bipolar system must balance against a rival because there is no other great power to “catch the buck.” A distant hegemon can safely stay out of any conflict in such regions because no state is powerful enough to conquer all of the states.

In multipolar systems states often buck-pass; states prefer to buck-pass rather than balance when confronted by a dangerous opponent. Buck-passing is most widespread when there is no potential hegemon to contend with and the threatened states do not share a common border. The more relative power the potential hegemon controls, the more likely it is that all of the threatened states in the system will forgo buck-passing and form a counterbalancing coalition. Also, for defensive realists, buck-passing becomes more likely among great powers in a multipolar system when the offense–defense balance favors the defense. Under this circumstance, more insulated or geographically distant great powers will tend to overestimate the ability of their frontline allies to withstand an initial attack, which in turn makes buck-passing a more attractive strategy for the former than balancing (Christensen 1997).

For Waltz and Mearsheimer external balancing is often slow and inefficient. If the regional great powers cannot contain the threat, the distant hegemon will balance against it. Thus, regional hegemons act as offshore balancers in other areas of the world and prefer to be “the balancer of the last resort.” Geography also helps identify the likely buck-passers and buck-catchers in multipolar systems: common borders promote balancing while barriers and buffers encourage buck-passing.

For Walt and other defensive realists, great powers rarely bandwagon. Only weak states with no great-power patron adopt such a risky strategy since there is no guarantee that the aggressor state will be satisfied rather than turning on and attacking the bandwagoners in the future (Ayoob 1989; David 1991). For defensive realists it is much safer to balance against rather than bandwagon with power or a threat. Schweller challenges the assumption that balancing behavior predominates among great powers. For Schweller, this assumption reflects what he calls the “status quo bias” which is prevalent among defensive realist arguments. Specifically, the bias assumes that “states are willing to pay high costs and take great risks to protect the values they possess, but will only pay a small price and take low risks to improve their position in the system” (1994:85). Only for satisfied countries can it be said that the primary goal is “to maintain their position in the system” (1994:86). Revisionist states want to increase their power and to improve their position in the international system. Calling for a “new order,” he argues that dissatisfied states are attracted to expanding revisionist powers. As already discussed, Schweller contends that jackal bandwagoning occurs when a powerful revisionist state or a coalition attracts opportunistic revisionist powers. The goal of such states is profit.

Differences among Offensive Realists

Offensive realists disagree on a number of important points. One debate is which distribution of power is most dangerous and exerts the greatest pressure for aggression, expansion, and conquest. For Mearsheimer, bipolar systems are the most stable and unbalanced multipolar systems are the least stable, the stopping power of water makes it impossible for a single state to become a global hegemon (hence there is no discussion of unipolar or hegemonic distributions), regional hegemons are defined by land armies, and oceans make it difficult for land powers to project their armed power over long distances necessary to create global hegemony.

For Gilpin and other hegemonic realists (Organski 1968; Modelski 1987; Thompson 1988), hegemonic systems where power is concentrated are the most stable distribution of power while systems approaching parity or where power is deconcentrated provoke expansion, conquest, and hegemonic war. For Gilpin, a global hegemon, which has historically been a naval power including the Netherlands, the UK, and the US, establishes and enforces the basic rules and rights that influence their own behavior and that of the lesser states in the system. Specifically, he claims that “the dominant economic (and military) powers in the modern era assumed the responsibility of organizing and defending the world market economy; they promoted free trade, provided investment capital, and supplied the international currency. In effect, they provided the public goods necessary for the functioning of efficient world markets because it was profitable for them to do so” (1981:139). Moreover, Gilpin recognizes that although control over an international system provides economic benefits to the dominant power or powers, domination also involves costs in manpower and material resources – what Paul Kennedy termed “imperial overstretch” (1987). The protection and related costs are not productive investments but instead constitute an economic drain on the economy of the dominant state. In time, there will be diminishing returns and increasing costs, which will eventually make it profitable for a potential hegemon to challenge the erstwhile hegemon for leadership over the international system.

Defensive Realism

The anarchical nature of the international system leads to different outcomes for offensive and defensive realist theorists. Offensive realists argue that anarchy compels states to maximize influence, to compete for power in a never-ending struggle for hegemony, and to engage in territorial expansion. For defensive realists, anarchy encourages states to adopt defensive, moderate, and restrained strategies (Jervis 1979; Van Evera 1984; 1999; Gaddis 1987; Walt 1987; Snyder 1991; Glaser 1994/5; Lynn-Jones 1995; Taliaferro 2000/01). For defensive realists, conflict is sometimes necessary such as in the case of aggressor states, when their security is threatened, when they are insecure, or when differences are irreconcilable (for a critique of defensive realism, see Vasquez 1997; Zakaria 1998; Legro and Moravcsik 1999; Mearsheimer 2001). Moreover, it is sometime difficult for satisfied states to identify each other (Jervis 1999). In general, defensive realists maintain that states seek to maximize security, preserve the existing distribution of power, are not inherently aggressive, and avoid relative losses due to shifts in their relative position and ranking (Grieco 1990). Kenneth Waltz argues that “In anarchy, security is the highest end. Only if survival is assured can states safely seek such other goals as tranquility, profit, and power” (1979:126). However, responding to realists who contend that states seek to maximize power and influence, he is clear that “the first concern of states is not to maximize power but to maintain their positions in the system” (1979:126).

For defensive realists there are four dangers from aggression, expansion, and conquest: First, attempts to achieve hegemony are self-defeating and can leave the state weaker and less secure because it provokes counterbalancing behavior and aggression tends to meet resistance. Defensive realists largely build on Waltz’s neorealist balance-of-power theory. Specifically, defensive realists begin with the supposition that balances of power recurrently form in the international system and that periods of sustained hegemony are not durable or stable. As Stephen Walt notes, “If balancing is more common than bandwagoning, then states are more secure because aggressors will face combined opposition. Status quo states should therefore avoid provoking countervailing coalitions by eschewing threatening foreign and defense policies” (1985:4). Moreover, for Walt, in addition to aggregate power, both offensive capability and offensive intentions will provoke counterbalancing behavior (“states with large offensive capabilities are more likely to provoke an alliance” and “states that appear aggressive are likely to provoke others to balance”; 1985:11–12). Thus, as Snyder reminds us, “the balance of power that arises out of international anarchy punishes aggressors, it does not reward it” (1991:11).

Second, conquest rarely pays. The cost of expansion usually exceeds the benefits and therefore expansion is often explained by non-systemic forces or domestic and unit-level pathologies. Third, the offense–defense military balance often favors defenders and the defense over the offensive. Finally, socialization and lessons from history teach states that expansion and the pursuit of hegemony are often misguided because they provoke counterbalancing rather than bandwagoning behavior.

One additional point is that in contrast to the assumption of offensive realism, states, and especially major powers (rather than small states), are fairly secure and much more so than any single individual (Bull 1977). As Robert Jervis notes, “one of the main reasons why international life is not more nasty, brutish, and short is that states are not as vulnerable as men are in a state of nature” (1978:172). While individuals can cease to exist fairly easily the same is not true for great powers. The consequence, as Jervis tells us, is that states that can afford to be cheated or that cannot be destroyed in a surprise attack can more easily trust others, can afford to wait for unambiguous signs of aggression, and do not need to engage in unbridled expansion for security (1978:172).

In sum, for defensive realists conquest is generally costly and the international system usually encourages moderation, the international system rarely encourages expansion and only under specific circumstances, domestic and unit-level pathologies often contribute to self-defeating behavior such as overextension and self-encirclement, and cooperation is possible among states and it is possible for status quo powers to signal their intentions to each other (though as Jervis notes there is the danger of misidentifying expansionist states as security-seekers). Finally, for Grieco and Waltz, concerns about relative gains (not “will both of us gain?” but “who will gain more?”), the fear of dependency and vulnerability, and cheating are the main barriers to cooperation among states. Offensive realists, as discussed above, contend that defensive realists exaggerate the restraint of the international system on the major states.

Conquest Rarely Pays

For defensive realists, conquest rarely pays. The reasons are manifold: aggression and military buildup will provoke counterbalancing alliances (Layne 2006b); modern nationalism makes conquest costly because it “spurs the defenders to fight harder” (Jervis 1978:195), makes it hard to subdue and manipulate people in defeated states, and repression will provoke massive popular resistance; modern information economies are difficult to subjugate, especially those that are built around information technologies and depend on openness and freedom of movement and transaction to function smoothly. Moreover, skilled labor may be more difficult to exploit. In addition, the nuclear revolution and second strike capability make it difficult for states to fight each other (Jervis 1990). Finally, control over politically hostile societies is expensive; the price of maintaining empire and especially the high levels of defense spending erode a great power’s economy; economic resistance and repression will reduce modern societies’ social surplus; and the gains from conquest are rarely additive (Kaysen 1990; Van Evera 1999). Thus, it is difficult to exploit conquered territories. Offensive realists counter by arguing that the gains from conquest are greater and the barriers to the formation of counterbalancing alliances, especially in (balanced) multipolar distributions, are higher than defensive realists recognize.

Security Dilemma

The security dilemma is one tragedy of anarchy for defensive realists. Many of the means that a state uses to increase its security will lower the security of other states, even among security-seeking powers. For John Herz (1951) and Robert Jervis (1976), one state’s attempt to increase its own security due to the anarchic nature of the international system can inadvertently threaten other states and make them less secure and thereby provoke them to augment their power. As Jervis tells us, “many of the means by which a state tries to increase its security decrease the security of others” (1978:169) or “an increase in one state’s security decreases the security of others” (1978:186). For defensive realists, motivated by security-seeking behavior, the outcome is an unintended hostility spiral among states that harbor no aggressive or revisionist intent. In fact, both states “would be satisfied with mutual security” (Jervis 1999). The end result is that the initiating state might undermine its security. As discussed below, “structural modifiers” such as geography and technology can ameliorate or exacerbate the security dilemma.

For Jervis and other defensive realists, there are several means to reduce the security dilemma: increasing the joint gains from cooperation; increasing the costs from non-cooperation; reducing the unilateral gains from the sucker’s payoff; and increasing the costs from mutual defection are among a few of the strategies (1978; 1985). Charles Glaser’s contingent realism adds to the discussion by arguing that under many conditions, rivals can achieve a higher level of security through coordination and cooperation rather than arms competition, aggression, and territorial expansion. For instance, he maintains that “when the risks of competition exceed the risks of cooperation, states should direct their self-help efforts toward achieving cooperation” (1994/5:60).

Status Quo States

For offensive realists, all states harbor revisionist intentions with hegemony as their ultimate goal. For defensive realists, it is possible for states that are satisfied with the status quo to signal their benign intent to each other and to identify each other. For Glaser, states that limit offensive capabilities through arms control, unilateral defense, and unilateral restraint, especially when the offense has the advantage, can communicate their benign intentions and motives to other states – that they are security-seeking states. Concomitantly, this can increase their security by preventing the security dilemma, hostility spirals, and arms races. Consenting to limit offensive capabilities when the defense has the advantage communicates less information about a state’s intentions and motives (Glaser 1994/5:68–9).

Structural Modifiers: Offense–Defense Military Balance and Geography

Defensive realism makes a number of refinements to Kenneth Waltz’s overly parsimonious structural theory of balance of power. For defensive realists, the international system itself is rarely sufficient to encourage states to seek to maximize power (Jervis 1999). Therefore, defensive realists emphasize the influence of the “fine-grained structure of power” or “structural modifiers” such as geography and technology and other factors (military doctrine, national social structure, diplomatic arrangements) in addition to the distribution of aggregate capabilities or the “gross structure of power” to explain expansion, aggression, and war (Snyder 1996:168–71; Van Evera 1999:7–8; Taliaferro 2000/01).

According to Van Evera (1999), when technology makes conquest easier: states are less secure and less likely to cooperate or engage in diplomacy; states cannot increase security without threatening others; there are greater incentives for preemption and to strike first and for “opportunistic expansion”; and strategies of security through expansion should be widespread even for status quo powers who must behave like aggressors in order to defend themselves against aggressors (one definition of offensive and defensive advantage is “when we say that the offense has the advantage, we simply mean that it is easier to destroy the other’s army and take its territory than it is to defend one’s own. When the defense has the advantage, it is easier to project and to hold than it is to move forward, destroy, and take” (Jervis 1978:187). For Glaser and Kaufmann the offense–defense balance is defined as “the ratio of the cost of the forces that the attacker requires to take territory to the cost of the defender’s forces”; 1998:46). For defensive realists, such situations of offensive advantage are a rare occurrence, especially given the second strike capability of most nuclear powers. Experts on ground warfare usually claim that the defender almost always enjoys a net advantage. Moreover, defensive realists hold that offensive dominance is often more perceptual than reality based (Van Evera 1998:6).

When technology makes conquest more difficult then the reverse holds (Van Evera 1999): states are more secure, have a more relaxed view, can wait for unambiguous signs of aggression (or until intentions and motives are clearer), can make themselves more secure without threatening other states, and status quo states can cooperate fairly easily and engage in diplomacy. For Jervis, defense dominance will “render international anarchy relatively unimportant” (1978:187). Thus, for defensive realists, when the defense is dominant states have little incentive to engage in territorial expansion. Moreover, “when the defensive has the advantage, status-quo states can make themselves more secure without gravely endangering others” (Jervis 1978:187)

One criticism of the offense–defense balance by offensive realists (and by others) is that it is often difficult to assess (Davis et al. 1998–99; Mearsheimer 2001). It is often unclear whether the offense or the defense has the advantage and whether it is possible to distinguish offensive from defensive strategies, whether a piece of military hardware is offensive or defensive often depends on the situation or is ambiguous, and under certain circumstances, even status quo powers will want offensive weapons (again, Mearsheimer assumes that all states have at least some offensive capabilities; Levy 1984; 1990/91; Lynn-Jones 1995; Glaser and Kaufmann 1998). For offensive realists, the consequence is that if there is any doubt about the offense–defense balance, states will assume a worst-case scenario and behave as though offensive strategies have the military advantage over defensive ones, thereby encouraging states to take advantage of windows of opportunity to expand, to aggregate power, and to block the rise of other states. As Walt notes, “If states cannot measure the offense–defense balance or distinguish between offensive and defensive capabilities, then security-seeking states cannot escape the security dilemma and cannot signal their peaceful intentions in a convincing manner” (2002:206).

A second structural modifier for defensive realists is geography. Similar to Mearsheimer’s view, geography can make aggression easier or more difficult. In land warfare, natural buffers and barriers (oceans, mountains, large rivers, deserts), size of territory, and difficult terrain aid the defender against superior numbers. For Mearsheimer, the stopping power of water makes it difficult for states to project their power over long distances. Furthermore, the loss-of-power gradient or the logistical burden of projecting power over a long distance tends to reduce the relative fighting power of the attacker.

Geography and technology can also affect the intensity and character of alliances and balancing behavior (Christensen and Snyder 1990). Where geography and technology are believed to favor the defender (defensive borders, large size, and strategic depth all protect against surprise attack (Jervis 1978), balancing behavior should be slower, involve more buck-passing, and be less intense than if the offensive is believed to have the advantage (Christensen and Snyder 1990). States with geographical defensive advantages will react more slowly and less intensively than other states to increases in an adversary’s power and will more often stress defensive military strategies. Where geographic factors reward offensive military strategies, states will be drawn to such doctrines whether or not they have status quo policies, and balancing behavior should tend to be quick and robust. Buck-passing should also occur less often, as Mearsheimer reminds us.

Socialization

For Waltz and for defensive realists, socialization to the norms of the system and learning lessons from history are important deterrents to expansion and aggression. According to Benjamin Frankel, “states are socialized into the system by emulating the practices of the most successful states in the system” (1996:xvii). As Walt notes, “if balancing is the norm and if states understand this tendency, aggression is discouraged because those who contemplate it will anticipate resistance” (1985:13). Similarly, Layne states that “one of history’s few incontestable lessons is that the pursuit of hegemony invariably is self-defeating” (2006a:6). Balance-of-power theory itself suggests that expanding hegemons will be opposed and stopped, and these lessons have been repeatedly demonstrated – though as offensive realists counter, not all states learn the lesson. Defensive realists reply that enough learning takes place to make violent, excessive, and expansionist policies the exception rather than the rule.

Domestic Pathologies

For defensive realists, the international system provides few incentives for expansion, it is usually difficult, and rarely profitable. Defensive realists see ambitious attempts to expand as self-defeating anomalies and the product of non-systemic or domestic political pathologies. Leaders might inflate threats to mobilize domestic resources (Christensen 1996), perceptions of the balance of power can affect state behavior (Leffler 1992; Wohlforth 1993; Christensen 1997), and aversion to loss can lead to risky diplomatic and foreign military intervention (Taliaferro 2004). According to Aaron Friedberg, states find it difficult to assess relative power which often lags behind shifts in the real distribution of power. Instead of steady, robust, and efficient balancing behavior (either through internal military buildup or counterbalancing alliances), a state’s adjustment is likely to be irregular, jerky, and to occur in bursts. The consequence is that a “nation whose leadership does not realize that its power is declining relative to that of another country will probably not feel compelled to enter into productive alliances with third parties” (1988:6).

According to Jack Snyder (1991), parochial groups (imperial and military interests) which disproportionately benefit from expansion but are too narrow to capture the state join with other pro-imperial interests to form a powerful logrolled coalition. Such logrolled coalitions have the greatest opportunity to control state policy where power is highly centralized – in cartelized political systems such as Germany and Japan in the 1930s (rather than democratic states such as Britain in the nineteenth and early twentieth century and the US during the Cold War where power is more diffused). The consequence is not expansion, but what defensive realists term overexpansion, or, more expansion than any single parochial group desired since each group will get the policy or program of expansion that it most prefers. However, multiple expansions can result in strategic over-commitment and self-encirclement. In the long run even the pro-imperial parochial groups are harmed by the counterproductive expansion.

For Charles Kupchan (1994), strategic culture or deeply embedded conceptions and notions of national security take root among elites and the public. Strategic culture can place severe constraints on the ability of elites to undertake strategic adjustment. By selling powerful strategic images to the polity which mold public opinion and reshape the roles and missions of the broader decision-making community, elites unwittingly entrap themselves in a strategic culture. This later prevents them from reorienting grand strategy and avoiding self-defeating behavior. The outcome is strategic exposure, self-encirclement, and overextension. In each instance the state is left pursing policies that jeopardize its primary security interests.

For Randall Schweller (2006), underbalancing runs contrary to the core prediction of structural realism. Namely, threatened states will balance against dangerous aggregations of power by forming alliances or building arms. Underbalancing occurs when the state does not balance or does so inefficiently in response to a dangerous aggressor, and the state’s efforts are essential to deter or defeat it. For Schweller, under-reactions to dangerous shifts in relative power occur for two reasons: actors’ preferences which are more influenced by domestic rather than international concerns do not create incentives to adopt a balancing policy, or the potential domestic political risks and costs of balancing are deemed too high. Thus, domestic constraints can prevent states from balancing in a timely and systematic manner in response to dangerous changes in relative power.

Conclusion and Policy Implications

According to offensive realism, security in the international system is scarce. Systemic imperatives of anarchy compel the major great powers to maximize their security through maximizing their relative power by expansionist foreign policies, taking advantage of opportunities to gain more power, and weakening potential challengers. For offensive realists, a state’s ultimate goal is hegemony or primacy because it is the best way to increase its odds of survival and it can never be certain of the relative power or future intentions of the other major powers. One consequence is that cooperation and trust among states are low. How a state will go about expansion will vary and offensive realism does not predict the same security strategy for every state, nor are states mindless expanders. However, offensive realists are clear that expansion often pays and that states are locked in a perennial struggle for power (including prestige, security, influence, and material capabilities).

For defensive realists, behavior beyond a moderate, incremental foreign policy is often unnecessary and counterproductive. The international system encourages moderate behavior and enough power; anything else must be explained at some other level of analysis. Defensive realists regard overexpansion, self-encirclement, and overextension as pathological and the product of domestic and unit-level variables (general staffs, perceptions, loss aversion, myths, domestic coalitions and cartels). For defensive realists, aggression is rare because states balance against aggressors and the offense–defense balance usually favors the defense thereby making conquest more difficult. Thus, great powers are often satisfied with the existing balance of power, rarely seek to change it through military force, security is abundant rather than scarce, and states have little incentive to seek additional power.

Offensive and defensive realism present opposing policy prescriptions and advocate grand strategies of offshore balancing, selective engagement, or primacy (for an overview of strategies, see Posen and Ross 1996/97). For offensive realists such as Mearsheimer, regional hegemons should pursue primacy in their locale and seek to block any peer rivals in other regions. For the US, this means that it will remain the regional hegemon in the western hemisphere and will act as an offshore balancer in Europe and Asia, though Germany, Russia, and China could emerge as potential regional hegemons, shifting both regions to unbalanced multipolar – the most unstable and war-provoking distribution. For Brooks and Wohlforth (2008) the current unipolar distribution of power is both stable and durable (Mastanduno 1997; Kaufman et al. 2007). The authors call for “systemic activism” by the US to use its unique window of opportunity to reshape the international system to reflect its long-term security interests. This behavior is possible because the US does not face the traditional external or systemic constraints that previous major powers encountered in bipolar and multipolar distributions (including the threat of counterbalancing by secondary states).

Defensive realists argue that the US should pursue either selective engagement or offshore balancing grand strategies rather than primacy. For Chris Layne, the danger of America’s current pursuit of predominance or “extraregional hegemony” is that it will provoke the emergence of old states and new powers to counterbalance the US, it encourages terrorists to target the US, the US will become entangled in overseas commitments, and finally it contributes to imperial overstretch which will erode US predominance (2006a:7). American power has not yet provoked hard balancing (Ikenberry 2002) but in contrast to Brooks and Wohlforth’s assessment, a number of defensive realists argue that it has provoked soft balancing or “tacit balancing short of formal alliances” (Paul 2004:3; for a critique of soft balancing, see Brooks and Wohlforth 2005). For Posen (2006), the concentration of American global power has led the European states, through the European Security and Defense Policy (ESDP), to move toward greater autonomy and to “cautiously” balance US power. For Robert Pape, American preponderance has led secondary states to “to engage in the early stages of balancing behavior against the United States” (2005:9), for T.V. Paul, the result is to form diplomatic coalitions against the United States (2005), and for Layne (2006b), the consequence has been leash-slipping by its allies.

One grand strategy for the US is selective engagement. According to Robert Art, “selective engagement is a strategy that aims to preserve America’s key alliances and its forward-based forces…it establishes priorities: it assures protection of America’s vital and highly important interests” (2003:9). Selective engagement has been criticized because it has not been very selective, with US forces still engaged in places like Europe despite the disappearance of the Soviet Union (Walt 2005). An alternative grand strategy, according to Walt and Layne, is for the US to act as an offshore balancer, intervening in vital locales when there are threats to American interests. The advantage of this strategy for Walt is that “it husbands the power upon which U.S. primacy rests and minimizes the fear that U.S. power provokes” (2005:223). One criticism of offshore balancing is that a less engaged US will “let the world become more dangerous and less prosperous” (Art 2003:222).